Module Redirect Id Based Urls

[adrian]

This module redirects ID based URLs to the full SEO friendly URL

 

Modules directory: http://modules.processwire.com/modules/process-redirect-ids/

Github: https://github.com/adrianbj/ProcessRedirectIds

 

Very simple module that alllows you to link to pages with their page ID in the URL. The ID can be in any location in the URL.

 

There is a config option to simply load the content to the ID based URL, rather than redirecting to the original PW url if you prefer. If you use this option, the module adds a canonical link to the page to help identify the ID based URL as a duplicate of the proper PW url so Google won't penalize you for duplicate content.

 

For example you could do any of the following, where 1058 is the ID of the page you want to load:

 

http://www.mysite.com/1058
http://www.mysite.com/1058/this-is-the-fancy-title
http://www.mysite.com/category1/category2/1058/any-text-you-want/
http://www.mysite.com/category1/category2/any-text-you-want-1058/  

Any of those will redirect to the proper URL, eg:

http://www.mysite.com/this-is-the-fancy-title/

At the moment for this to work, the page ID segment of the URL must be the only segment that is all numerical, otherwise, anything goes.

 

The reason for this module is mostly as a way to have a very simple shortlink to any page on the site, but it also allows for structuring URLs however you want with any content you want. I guess this could be abused and potentially bad for SEO!

 

I am not sure if this is a good way to do this, so I would definitely appreciate any feedback.

[diogo]

I thought of this already, but gave up on the idea because it allows anyone to start trying random numbers and get pages that you don't want accessible to everyone. One way of avoiding this would vbe to have a global checkbox field that you would have to check to activate the shortlink to that page.

[adrian]

Hey diogo - I agree that could be an issue, but surely if pages are viewable then they will be indexed by search engines anyway. The module does check to make sure the page is not an admin page and that it is viewable for the user. Maybe that's not enough?

I'd be happy to add the checkbox idea though if you think that would help. Perhaps a "disable if checked" approach, or maybe a module config setting to change the behavior of the page specific checkbox - either check to enable, or check to disable, so there is the option to go either way.

Another option might be to require it to be enabled on a per template basis - an ASM field in the module config would allow you to specify templates that would work.

Edited September 27, 2013 by adrian

[diogo]

The page is not indexed if there's nothing linking to it. I would have the links disabled by default at least and the checkbox to enabled them, and maybe the possibility of enabling by template (as you said) or by parent because it makes all the sense to enable all descendants of a page (articles page for instance). Like this, you could enable the links to all the site just by enabling the descendants of home. Or am I complicating?

[alan]

Hey Adrian, thanks for the Module, if I read that right, it looks like functionality I remember from Textpattern? If so then it was something I was trying to do (very poorly) in this thread, linked just in case it's any help (can't believe it will be tho )—actually that thread title is misleading, the post I was referring to was this one on that thread.

Edited September 29, 2013 by alanfluff

[Marty Walker]

IF this module ends up being is a goer (security implications sorted n'all) would it be possible to have some way in the admin of letting the user copy the short url to that page. I can see where this would come in handy for a site or two.

[adrian]

@diogo - I have added support for limiting this functionality to only those selected templates (in the module config settings). I was also trying to add support for by page, but have come across a very strange thing - I can't seem to get an array of pages using wire('pages'). I get:

Trying to get property of non-object in /share/www/usaus-h2o-dev/site/modules/ProcessRedirectIds/ProcessRedirectIds.module on line 179

I have commented out this section of the code for the moment, but if you have any ideas why this isn't working, I'd love to hear them

@alanfluff - thanks for that link - I think this module will take care of what you are looking to do.

@Marty Walker - I have added a new tab to page edit screen in the admin called ShortLinks that lists a few key links for that page that the user might be interested in. The last example I give on that tab makes me wonder whether this module should actually be renamed to something that mentions permalinks.

I am not sure where to go with this now - I will definitely be using this for an upcoming site, but I would really like some feedback from some other PW gurus, especially Ryan, as to whether this should be a released module and whether you guys think this idea of a permalink like:

http://www.mysite.com/1234/articles/my-article-title/

is a good idea to be promoting or not. I know that Page Path History does an amazing job already of handling changes to page names and parents. So would the idea of using this module to create this style of permalink be redundant, or just bad practice?

Thanks for any feedback.

[alan]

@adrian This module is exactly what I was looking at doing before, but done better, so thanks again.

I've not read up on Page Path History but I see your question re that or this module. I like the idea of this module, my only comment would be that perhaps the approach the BBC take might be better, it's the same as yours but instead of

http://www.mysite.com/1234/articles/my-article-title/ 

they use

http://www.mysite.com/articles/my-article-title-1234

I assume as the result may be slightly better SEO-wise.

Edited September 30, 2013 by alanfluff

[diogo]

I can't seem to get an array of pages using wire('pages'). I get:

What's the idea on these lines?

foreach(wire('pages') as $p) {
            $fieldEnabledPages->addOption($p->name);
        }

Grab all the pages?

funny, wire("pages") or even wire("pages")->find() also don't return anything to me (the same happens with $pages on templates, by the way), but wire("pages")->find("sort=sort") does. Maybe you can try that.

try:

foreach(wire("pages")->find("sort=sort") as $p) {
            $fieldEnabledPages->addOption($p->name);
        }

Anyway, I'm not sure if it is a good idea to iterate all the pages of the site in a module because it might cause problem on bigger sites.

[adrian]

@alan - I have added support for your suggested URL structure. Please test when you get a chance and let me know how it goes for you.

@diogo - thank you for the testing. The ->find("sort=sort") works for me too - any ideas why it doesn't work as normal?

Good point about iterating all those pages on a large site - that definitely won't work well at all. I have changed config setting to a InputfieldPageListSelectMultiple and it seems to work well now.

So now it is possible to limit the redirection of these links by template and or page/page parent.

Any other thoughts/ideas? 

[adrian]

I asked Ryan about this module and he is happy with it going in the modules directory, so I have submitted it:

http://modules.processwire.com/modules/process-redirect-ids/ 

It is fine to have it run on all your pages/templates (the default), but don't forget to check out the module config settings where you can limit this. Remember it is only called by pages that return a 404 so I don't believe there should be any overhead on normal page loads.

So please feel free to start using it and let me know if you have any improvements you'd like to see.

[alan]

Sorry I didn't get to test it Adrian, I'm sure it would have been a redundant test and I'll be trying this out next site I build, congrats on the Module and thanks!

[webweaver]

This module accomplishes exactly what I was looking for today. Thank you very much Adrian for making this available.

[hellomoto]

IF this module ends up being is a goer (security implications sorted n'all) would it be possible to have some way in the admin of letting the user copy the short url to that page. I can see where this would come in handy for a site or two.

Is this really such a security risk..? ...that people could guess IDs? It's not about whether those pages are indexed by Google; the same risk definitely applies to the default slug-based permalinks -- like Adrian said, he'd assume they would already be publicly inaccessible from the server-side, me too; it's no more secure than not doing so and leaving the regular slugs (unless you have yours or your titles randomly generated as complex strings). Or am I missing something?

Anyway I can't wait to try this out, thanks much.

[adrian]

@hellomoto,

I don't really think there is a security risk, especially given that you can limit this module's actions to specific templates and pages if you wish.

Anyway, hope it works as you expect!

[hellomoto]

Hey! Ok so it doesn't quite work as I expected, though it does work great... but what I'm looking to do is actually use the IDs for pages of one of my templates. 

I have it set up to Redirect (default), Rewrite Links checked, Rewrite Format = /sales/listings/{$page->make->name}/{$page->model->name}/{$page->id}/

/sales/listings/ is the path preceding each listing anyway, which is the template I have the rewrite enabled for.

So it works in my listings index on the frontend, but then those links redirect to their default URLs. I thought this might be accomplished with Load selected for Redirect Type instead of Redirect, but that gives me this error:

Compile Error: Cannot redeclare browserTitle() (previously declared in /.../site/templates/includes/functions.php:5) (line 10 of /.../site/templates/includes/functions.php)

Lines 5 and 10 are the beginning and ending of that function... So I don't know why that would be happening...

Update: I changed my header to include_once the functions. This is fixed

But what I really want is to redirect the default URL for each listing to the one I design. Like from the backend, the links still go to /sales/listings/{$page->name}/ when I want them to be what I set too, or at least redirect to what I set. I want the URL I set to be the absolute one. Anyone know how? Thanks.

[adrian]

The Load option should do exactly what you are looking for. 

You shouldn't have the /sales/listings at the beginning of the rewrite format - just use:

/{$page->make->name}/{$page->model->name}/{$page->id}/

I haven't tested thoroughly with adding extra path segments like you are doing, but it seems to be working fine here. Typically I just put dashes between the components.

Glad you sorted out the include_once issue

To get the View links when editing a page to also be rewritten, make this change to line 115:

     if($page->id!=1 && $page->template != 'admin') {

to:

     if($page->id!=1) {

I did a quick test and I don't think there are any issues with this, but would appreciate you testing it before I make it live.

[gmclelland]

I just tried out the module and I can confirm that it works just like it says.
 
I'm new to Processwire, so forgive me if what I'm suggesting is to much.  I come from the Drupal land.  In Drupal, all pages have a id called an nid(node id).  With core Drupal you can visit any page with /node/10003 and it will redirect you to that page.  When links from the editor are inserted into a textarea, they look like /node/10003.  When the page is viewed, the https://www.drupal.org/project/pathologic module converts those types of links into their aliased name(like /my-cool-page).

I believe Concrete5 does the same thing with it's core page link abstractor.
 
Couple of things/possible suggestions:
Is there a way to rewrite links that are inserted via the wysiwyg editor into a textarea?  For instance, clicking on the link icon in ckeditor opens a modal dialog where you can select a page, but when you select a page it inserts something like /parent/child-page instead of /1002.  If it just inserted /1002 then the editors wouldn't need to look at the short links tab because this module would automatically handle everything.
 
Maybe there can be a way to use a text formatter module to convert links like /1002 to /parent/child-page when outputting the page like https://www.drupal.org/project/pathologic?
 
Drupal also adds the following to the page to assist with SEO:
<link rel="canonical" href="https://www.mywebsite.com/my-cool-page" />
<link rel="shortlink" href="https://www.mywebsite.com/node/4796" />

Processwire looks really good, but I can't imagine not having some kind of way to handle links breaking when the site tree changes.  For big departmental or documentation websites, this happens all the time.  P.S. I have seen the other two modules page-link-abstractor and page-path-history.

Hope that helps, thanks for sharing your module I think it is definitely moving in the right direction!

[gmclelland]

Maybe it would be possible to add another option to the ProcessPageEditLink module? something like "Absolute page id"? that would insert links like /1002

[adrian]

Hi @gmclelland - to be honest I don't really use this module for replacement of the main links - I only use it as a means of providing a shorter URL if required.

I actually think that the Page Path History module does an excellent job of preventing broken links. I have never actually been a fan of Drupal's /node/nnnn/ links, but in previous custom systems I have built, I have included the page ID as part of the URL, like: mydomain.com/nnnn-title-of-my-page/ which this module allows.

As for changing the insert link - this should already work - at least it does for me:

Sorry if I am missing your point - in a bit of a hurry!

PS Did you select the "Rewrite Links" in the module settings?

[gmclelland]

Oh... That's what that setting does.    Yes, it works for me.  Thanks for pointing me in the right direction.  The only thing missing now is maybe some kind of text formatter module that would convert /1002 type links to /my-cool-page?

I've only used Page Path History once a while back, but it seems like Page Path History module is just covering up the problem by creating a ton of redirects without fixing the source of the problem which is the content is linking to the wrong page?  In Drupal, your not supposed to see the node/1002 links on the frontend.  Instead, you see the node's alias which would be something like /my-cool-page.

Their is a https://www.drupal.org/project/redirect module that is used to store changes of the node's alias.  For example, changing /my-cool-page to /my-cooler-page, but that is really only useful for when other websites link to the node's alias.  Internal links in text areas are always like node/1002 and then transformed on output into /my-cool-page.

As far as I know the systems that do this are modx, concrete5, Drupal.

Anyways, thanks again for responding.  I appreciate your help.

[Mike Rockett]

As far as I'm concerned, redirects are pretty quick, and so using normal page links in CKEditor along with PagePathHistory should be more than sufficient for managing this. A Textformatter module would also indeed solve the problem, but there isn't one as yet.

Also not a fan of ID-based URIs - I just don't see the purpose...

[adrian]

I have seen several drupal sites where the link in the html is still the /node/nnnn - does that mean they don't have it configured correctly or is this just something that is unavoidable in some cases, in which case that seems worse from an SEO point of view to me than just redirecting outdated links with PagePathHistory keeping in mind that there won't likely be that many changes - seems better than redirecting every /node/ link if it isn't being rewritten properly before the HTML is generated.

However, a textformatter to rewrite ID links made by this module wouldn't be hard to put together - in fact I bet you could steal some code from the Link Abstractor module - which btw is now deprecated.

If go to tackle this and need some help, please let us know and we'll help you get it working.

[gmclelland]

Thanks adrian for the help.

I have seen several drupal sites where the link in the html is still the /node/nnnn - does that mean they don't have it configured correctly 

Yes, that can happen if you don't set any page aliases with the https://www.drupal.org/project/pathauto module.  The Pathauto module is used on pretty much every Drupal site.

[Orkun]

Hi @adrian

Thanks for this great Module!

Is it possible to use your module with somas Multisite module? I tried it with somas latest module version and it didn't worked for me.

https://github.com/somatonic/Multisite/tree/dev2

Or should it already work? Have you tested with it? It would be very great!

 

Greetings Nukro